Closed jim22k closed 2 years ago
Allows conversion of dtype for sparse tensor. The layout must remain unchanged.
This is written in MLIR, not a new function in SparseUtils.cpp. So the casting is based on whatever LLVM does.
Allows conversion of dtype for sparse tensor. The layout must remain unchanged.
This is written in MLIR, not a new function in SparseUtils.cpp. So the casting is based on whatever LLVM does.